Reconstructive Classification for Age and Gender Identification in Social Networks

Grouping users in social networks is an important process that improves matching and recommendation activities in social networks. The data mining methods of clustering can be used in grouping the users in social networks. However, the existing general purpose clustering algorithms perform poorly on the social network data due to the special nature of users' data in social networks. One main reason is the constraints that need to be considered in grouping users in social networks. Another reason is the need of capturing large amount of information about users which imposes computational complexity to an algorithm. In this paper, we propose a scalable and effective constraint-based clustering algorithm based on a global similarity measure that takes into consideration the users' constraints and their importance in social networks. Each constraint's importance is calculated based on the occurrence of this constraint in the dataset. Performance of the algorithm is demonstrated on a dataset obtained from an online dating website using internal and external evaluation measures. Results show that the proposed algorithm is able to increases the accuracy of matching users in social networks by 10% in comparison to other algorithms.

The properties of singular value decomposition (SVD) are used to implement an SVD spatial domain pseudoinverse restoration filter. This type of filter is attractive for poor imaging conditions (low spatial resolution, high image noise) and is thus appealing for nuclear medicine images. The method might offer some advantages over more traditional frequency domain filter techniques since the restoration is performed on a local rather than global basis. High-contrast thyroid phantom images collected at different count densities and low-contrast liver phantom images were processed with the SVD filter. Restored images yielded improved spatial resolution, lesion contrast, and signal-to-noise ratio. The SVD pseudoinverse restoration filter implemented as an interactive process permits the operator to terminate filtering at a stage where the visually "best" image is obtained compared to the original data. Processed images suggest that the technique may have potential for improving lesion detection in nuclear medicine.

Nowadays, the user linkage or link prediction task is usually based on user profile or some sensitive data (i.e., name, gender, ID, life or health insurance accounts, etc.). With the enhancement of laws and regulations, the difficulty of personal sensitive data acquisition is increasing. Moreover, the abnormal users called online water army often camouflage themselves to achieve specific goals. They often register false user information such as name, gender, age, etc. To protect privacy and satisfy the needs of camouflage, users and ISPs often hide those sensitive data (i.e., user profile). In this paper, we want to link same user in multiple social networks, which is formally defined as ULASN (User Linkage across Anonymized Social Networks) problem. ULASN is very challenging to address due to (1) the lack of enough ground-truth to build models and obtain accurate prediction results, (2) the studied networks are anonymized, where no user profile or sensitive data is available, and (3) the need of scalable algorithms for user linkage task in large-scale social nateworks, and (4) users in social network are interrelated. To resolve these challenges, a noval user linkage framework based on social structures called ULA is proposed in this paper. ULA tackles these problems by considering massive, low-quality and interrelated user information. It uses few ground-truth to partition users into blocks, which reduces the size of candidates. By extending Fellegi-Sunter methods, our proposed algorithm can handle social network similarity complying to continuous distributions. A probabilistic generative model is proposed and solved by EM algorithm. Simultaneously, missing value problem can also solved when we use EM algorithm to learning parameters. Extensive experiments conducted on two real-world social networks demonstrate that ULA can perform very well in solving ULASN problem.

The security of information exchange is very important on the network. Authentication and information hiding have also become important issues. Information hiding techniques are acquiring an increasing importance due to the widespread diffusion of multimedia contents. The aim of this chapter is to focus on the Singular Value Decomposition (SVD) transform, with the aim of providing an exhaustive overview on those steganography, image cryptography and watermarking techniques leveraging on the important properties of such a transform. Despite the attention it has received in the last years, SVD in image processing and security is still in its infancy. Many SVD characteristics are still unutilized in image processing. In this chapter the author tries to highlight the basic properties of SVD and some of their applications in the field of security to encourage researchers to discover more about SVD properties which are not yet utilized.

We propose a technique for video shot segmentation and classification based on singular value decomposition (SVD). For the input video sequence, we create a feature-frame matrix A, and perform the SVD on it. From this SVD, we are able to not only derive the refined feature space to better segment the video sequence along the time axis, but also define metrics to enable classifications of the detected video shots. Using these SVD properties, we achieve the two goals of accurate video shot segmentation, and visual content-based shot classification at the same time.

Evaluation Of Recommendation System On A Movie Dataset

In this paper, two new properties of singular value decomposition (SVD) on images are proved. The first property demonstrates the quantitative relationship between singular values and power spectrum. The second one proves that under the condition of losing equal power spectrum, the square-error of the reconstructed image is much smaller when we reduce all singular values proportionally instead of neglect the smaller ones. Based on the two properties, a new data-hiding scheme is proposed. It performs well as for robustness, for it satisfies power-spectrum condition (PSC), and PSC-compliant watermarks are proven to be most robust. Besides, the proposed scheme has a good performance as for capacity and adaptability.

Block-wise reversible data hiding in encrypted domain using SVD

The internet presents a vast labyrinth of information and choices, overwhelming users while service providers strive to predict user preferences accurately. Traditional recommender systems, although utilizing extensive data, often lack the social context necessary for precise recommendations, leading to issues like cold-start and data sparsity. Social network-based recommender systems have been proposed to address these limitations by leveraging the decision-making processes influenced by social connections. Recent advancements in this domain, such as the Diffnet and Diffnet++ models, have shown improved prediction accuracy through deep social influence propagation models. However, their performance diminishes with higher diffusion depths. In this work, we propose enhancements to the Diffnet and Diffnet++ models by incorporating the homophily effect in the heterogeneous graph of the diffusion layer and implementing a weighted embedding method in the prediction layer. Using data from Yelp and Flickr, our improved model integrates higher-order social structures and interest network information, demonstrating superior performance on metrics such as Normalized Discounted Cumulative Gain (NDCG) and Hit Ratio (HR) compared to existing baselines. Our approach effectively integrates social network data and user similarity networks, resulting in more accurate and reliable recommendations for online platforms where social networking is a key feature.

The use of social media, such as social networking sites continues to grow in everyday life, along with their use in the workplace. So, a concern raised about how the personal use of social networking sites like Facebook, WhatsApp, Instagram, etc., at work affects employee job performance. Moreover, the purpose of this study is to explore the relationship between social networking usage and employee job performance through the mediating role of work–life balance. The quantitative methodology was employed in this study. Data was collected using a self-administered questionnaire adapted from a previously validated survey. The target population consisted of Gen-Y employees of the corporate sector in Chandigarh, Delhi, and Haryana (UTs/states) of India. Judgemental sampling was applied to collect data from 224 employees of the corporate sector. The study results indicated that time spent, frequency of using social networking sites has significant relationship with job performance and work–life balance of employees while work–life balance and employee job performance are significantly positively related. However, social networking use has a significant negative relationship with job performance and work–life balance. Additionally, work–life balance significantly mediates the relationship between social networking usage and job performance. Therefore, the study concluded that employees struggle to set boundaries since they cannot distinguish between work and family life due to the rise in technology and social media use. Therefore, managers and HR professionals need to consider the negative aspect of online social networking at work. Further, the implications for managers and organizations are discussed.

The appearance of social networks has been one of the most important events in the last decade. A social network is a network of interaction and communication whose nodes are its members and its edges are the relation between them. An outstanding concept in social networks is the concept of social influence which is of great significance. Social influence is the behavioral transformation in a person, caused by his/her relationship with other people, organizations and the network. Today, the concept of social influence is used in viral marketing. The aim of viral marketing is to find a small subset of influential users in the social networks for marketing a product. The aim of this paper is to offer an algorithm in order to choose expert and influential users in social networks. This algorithm uses leader discovery method in social networks for choosing expert and influential users. This method chooses those users as the expert and influential leaders of the social networks, who are influential among the other members and also those users that have enough specialty and knowledge about the marketing product. Consequently, the process of selecting expert and influential users for the algorithm would be quick and also the selected leaders are influential and expert enough to do marketing and advertising in the social network.

The purpose of research was to study the existing methods of determining the degree of cohesion of two users of social network, identifying their shortcomings and developing a new method. The research identified shortcomings of existing methods and proposed a new method for assessing the degree of cohesion of social network profiles based on open data from a social network. Under the degree of cohesion of users’ profiles is understood the probability of communication (interaction) of profile owners in real life, it is calculated for two users of the social network and expressed in percent. The work of the method is demonstrated on the example of the social network “In contact”. This method includes the sequence of the following stages: the first stage is data collection about users of the social network with API and the formation of tuples of users’ profile characteristics. A tuple of characteristics of social network profiles is the data, collected for each user, stored in a structured form. The next step is the analysis of the collected information. For each characteristic of the tuple of profiles, i.e. the possible element of interaction of users in the social network, the coefficient of cohesion by the characteristic is calculated. In addition, for each feature, its informativeness is calculated, i.e. how important is this or that feature in this social network. At the final stage, the results are generated, using the formula for the probability of communication between two users, derived during the investigation. Obtained as a result of the application of the method, the probability of communication between two users can be used to optimize the activities of the operative-search services and special bodies. In addition, the received degree of cohesion of two users can be interpreted as the probability of a channel of information leakage between them. The role of the user of the method can be any private or state organization that cares about the security of corporate data and commercial secrets, the operative-search service, as well as an organization that investigates cybercrimes and information security incidents.

Identification of influential users by neighbors in online social networks

Identifying the influential spreaders is an important issue in understanding and controlling the spreading processes in social networks. The key to the influential spreader identification problem is how to evaluate the spreading ability of the nodes. Centralities such as degree, semi-local centrality, betweenness, closeness, k-shell are usually used as evaluation metric. However, we observe that these centralities are sensitive not only to the spreading probability, but also to the network structure. We are not sure which centrality is efficient when we face a new network. In this paper, we propose a robust method named Adjustable multi-Hops Spreading (AHS). In AHS, we refine the influence of a node into direct influence and indirect influence, and then integrate them with a adjustable parameter. The experimental results in both real social networks and artificial networks show that AHS outperforms other centralities in effectiveness, robustness and the distinguish ability.

In this article, we have improved the quantitative inversion performance of the cross-correlated contrast source inversion (CC-CSI) method by incorporating the subspace optimization strategy. The proposed method is called the cross-correlated subspace optimization method (CC-SOM). Meanwhile, multifrequency data are used to improve the inversion performance of high-contrast scatterers, where the L-curve method is introduced to select the regularization parameters of each frequency point without relying on experience. Finally, a fast algorithm is implemented by using the property of singular value decomposition (SVD) to simplify the large-scale matrix, and the fast Fourier transform (FFT) to accelerate the calculation. Synthetic and experimental inversion results demonstrate that both CC-SOM and CC-CSI show better robustness than SOM. In comparison to CC-CSI, CC-SOM is superior in terms of inversion accuracy when the regularization parameters have been appropriately selected. However, these advantages come at the cost of higher computational complexity.